Kafka學習之路 （五）Kafka在zookeeper中的存儲

當kafka啓動的時候，就會向zookeeper裏面註冊一些信息，這些數據也稱爲Kafka的元數據信息。

一、Kafka在zookeeper中存儲結構圖

二、分析

根目錄下的結構

服務端開啓的情況下，進入客戶端的命令：{zookeeper目錄}/bin/zkCli.sh

# {zookeeper目錄}/bin/zkCli.sh
[zk: localhost:2181(CONNECTED) 1] ls /
[cluster, controller_epoch, controller, brokers, zookeeper, admin, isr_change_notification, consumers, log_dir_event_notification, latest_producer_id_block, config]

2.1　topic註冊信息

/brokers/topics/[topic] :

存儲某個topic的partitions所有分配信息

[zk: localhost:2181(CONNECTED) 1] get /brokers/topics/topic2

Schema:
{
    "version": "版本編號目前固定爲數字1",
    "partitions": {
        "partitionId編號": [
            同步副本組brokerId列表
        ],
        "partitionId編號": [
            同步副本組brokerId列表
        ],
        .......
    }
}
Example:
{
"version": 1,
"partitions": {
"2": [1, 2, 3],
"1": [0, 1, 2],
"0": [3, 0, 1],
}
}

2.2　partition狀態信息

/brokers/topics/[topic]/partitions/[0...N] 其中[0..N]表示partition索引號

[zk: localhost:2181(CONNECTED) 2] ls /brokers/topics/[topic]/partitions
[0, 1, 2]
[zk: localhost:2181(CONNECTED) 3] /brokers/topics/[topic]/partitions/[partitionId]/state

Schema:
{
"controller_epoch": 表示kafka集羣中的中央控制器選舉次數,
"leader": 表示該partition選舉leader的brokerId,
"version": 版本編號默認爲1,
"leader_epoch": 該partition leader選舉次數,
"isr": [同步副本組brokerId列表]
}

Example:
{
"controller_epoch": 1,
"leader": 3,
"version": 1,
"leader_epoch": 0,
"isr": [3, 0, 1]
}

2.3　Broker註冊信息

/brokers/ids/[0...N]

[zk: localhost:2181(CONNECTED) 4] ls /brokers/ids
[0, 1, 2, 3]
[zk: localhost:2181(CONNECTED) 5] get /brokers/ids/0

每個broker的配置文件中都需要指定一個數字類型的id(全局不可重複),此節點爲臨時znode(EPHEMERAL)

Schema:
{
"jmx_port": jmx端口號,
"timestamp": kafka broker初始啓動時的時間戳,
"host": 主機名或ip地址,
"version": 版本編號默認爲1,
"port": kafka broker的服務端端口號,由server.properties中參數port確定
}

Example:
{
"jmx_port": -1,
"timestamp":"1525741823119"
"version": 1,
"host": "hadoop1",
"port": 9092
}

2.4　Controller epoch

[zk: localhost:2181(CONNECTED) 6] get /controller_epoch
1

此值爲一個數字,kafka集羣中第一個broker第一次啓動時爲1，以後只要集羣中center controller中央控制器所在broker變更或掛掉，就會重新選舉新的center controller，每次center controller變更controller_epoch值就會 + 1;

2.5　Controller註冊信息

存儲center controller中央控制器所在kafka broker的信息

[zk: localhost:2181(CONNECTED) 7] get /controller

Schema:
{
"version": 版本編號默認爲1,
"brokerid": kafka集羣中broker唯一編號,
"timestamp": kafka broker中央控制器變更時的時間戳
}

Example:
{
"version": 1,
"brokerid": 0,
"timestamp": "1525741822769"
}

2.6　補充Consumer and Consumer group

① 每個consumer客戶端被創建時,會向zookeeper註冊自己的信息;
② 此作用主要是爲了"負載均衡"。
③ 同一個Consumer Group中的Consumers，Kafka將相應Topic中的每個消息只發送給其中一個Consumer。
④ Consumer Group中的每個Consumer讀取Topic的一個或多個Partitions，並且是唯一的Consumer。
⑤ 一個Consumer group的多個consumer的所有線程依次有序地消費一個topic的所有partitions,如果Consumer group中所有consumer總線程大於partitions數量，則會出現空閒情況。

舉例說明：

kafka集羣中創建一個topic爲report-log 4 partitions 索引編號爲0,1,2,3

假如有目前有三個消費者node：注意 --> 一個consumer中一個消費線程可以消費一個或多個partition

如果每個consumer創建一個consumer thread線程,各個node消費情況如下，node1消費索引編號爲0,1分區，node2費索引編號爲2,node3費索引編號爲3

如果每個consumer創建2個consumer thread線程，各個node消費情況如下(是從consumer node先後啓動狀態來確定的)，node1消費索引編號爲0,1分區；node2費索引編號爲2,3；node3爲空閒狀態

總結：

從以上可知，Consumer Group中各個consumer是根據先後啓動的順序有序消費一個topic的所有partitions的。

如果Consumer Group中所有consumer的總線程數大於partitions數量，則可能consumer thread或consumer會出現空閒狀態。

2.7　Consumer均衡算法

當一個group中,有consumer加入或者離開時,會觸發partitions均衡.均衡的最終目的,是提升topic的併發消費能力。

1) 假如topic1,具有如下partitions: P0,P1,P2,P3
2) 加入group中,有如下consumer: C0,C1
3) 首先根據partition索引號對partitions排序: P0,P1,P2,P3
4) 根據(consumer.id + '-'+ thread序號)排序: C0,C1
5) 計算倍數: M = [P0,P1,P2,P3].size / [C0,C1].size,本例值M=2(向上取整)
6) 然後依次分配partitions: C0 = [P0,P1],C1=[P2,P3],即Ci = [P(i M),P((i + 1) M -1)]

2.8　Consumer註冊信息

每個consumer都有一個唯一的ID(consumerId可以通過配置文件指定,也可以由系統生成),此id用來標記消費者信息。

/consumers/[groupId]/ids/[consumerIdString]

是一個臨時的znode,此節點的值爲請看consumerIdString產生規則,即表示此consumer目前所消費的topic + partitions列表。

consumerId產生規則：

StringconsumerUuid = null;
    if(config.consumerId!=null && config.consumerId)
      consumerUuid = consumerId;
    else {
      String uuid = UUID.randomUUID()
      consumerUuid = "%s-%d-%s".format(
        InetAddress.getLocalHost.getHostName, System.currentTimeMillis,
        uuid.getMostSignificantBits().toHexString.substring(0,8));

     }
     String consumerIdString = config.groupId + "_" + consumerUuid;

[zk: localhost:2181(CONNECTED) 8] ls /consumers
[zk: localhost:2181(CONNECTED) 9] ls /consumers/console-consumer-2034/ids
[zk: localhost:2181(CONNECTED) 10] get /consumers/console-consumer-2034/ids/console-consumer-2034_kafka1-8454138431234-7a59ee48

Schema:
{
"version": 版本編號默認爲1,
"subscription": { //訂閱topic列表
"topic名稱": consumer中topic消費者線程數
},
"pattern": "static",
"timestamp": "consumer啓動時的時間戳"
}

Example:
{
"version": 1,
"subscription": {
"topic2": 1
},
"pattern": "white_list",
"timestamp": "1525747915336"
}

2.9　Consumer owner

/consumers/[groupId]/owners/[topic]/[partitionId] -> consumerIdString + threadId索引編號

a) 首先進行"Consumer Id註冊"。
b) 然後在"Consumer id 註冊"節點下注冊一個watch用來監聽當前group中其他consumer的"退出"和"加入";只要此znode path下節點列表變更,都會觸發此group下consumer的負載均衡.(比如一個consumer失效,那麼其他consumer接管partitions)。
c) 在"Broker id 註冊"節點下,註冊一個watch用來監聽broker的存活情況;如果broker列表變更,將會觸發所有的groups下的consumer重新balance。

[zk: localhost:2181(CONNECTED) 11] ls /consumers
[console-consumer-2034]
[zk: localhost:2181(CONNECTED) 12] ls /consumers/console-consumer-2034/owners
[topic2]
[zk: localhost:2181(CONNECTED) 13] ls /consumers/console-consumer-2034/owners/topic2
[0, 1, 2]
[zk: localhost:2181(CONNECTED) 14] get /consumers/console-consumer-2034/owners/topic2/0

2.10　Consumer offset

/consumers/[groupId]/offsets/[topic]/[partitionId] -> long (offset)

用來跟蹤每個consumer目前所消費的partition中最大的offset。

此znode爲持久節點,可以看出offset跟group_id有關,以表明當消費者組(consumer group)中一個消費者失效，

重新觸發balance,其他consumer可以繼續消費。

[zk: localhost:2181(CONNECTED) 15] get /consumers/console-consumer-2034/offsets/topic2/0

2.11　Re-assign partitions

[zk: localhost:2181(CONNECTED) 16] get /admin/reassign_partitions

{
   "fields":[
      {
         "name":"version",
         "type":"int",
         "doc":"version id"
      },
      {
         "name":"partitions",
         "type":{
            "type":"array",
            "items":{
               "fields":[
                  {
                     "name":"topic",
                     "type":"string",
                     "doc":"topic of the partition to be reassigned"
                  },
                  {
                     "name":"partition",
                     "type":"int",
                     "doc":"the partition to be reassigned"
                  },
                  {
                     "name":"replicas",
                     "type":"array",
                     "items":"int",
                     "doc":"a list of replica ids"
                  }
               ],
            }
            "doc":"an array of partitions to be reassigned to new replicas"
         }
      }
   ]
}

Example:
{
  "version": 1,
  "partitions":
     [
        {
            "topic": "Foo",
            "partition": 1,
            "replicas": [0, 1, 3]
        }
     ]            
}

2.12　Preferred replication election

[zk: localhost:2181(CONNECTED) 17] get /admin/preferred_replica_election

{
   "fields":[
      {
         "name":"version",
         "type":"int",
         "doc":"version id"
      },
      {
         "name":"partitions",
         "type":{
            "type":"array",
            "items":{
               "fields":[
                  {
                     "name":"topic",
                     "type":"string",
                     "doc":"topic of the partition for which preferred replica election should be triggered"
                  },
                  {
                     "name":"partition",
                     "type":"int",
                     "doc":"the partition for which preferred replica election should be triggered"
                  }
               ],
            }
            "doc":"an array of partitions for which preferred replica election should be triggered"
         }
      }
   ]
}

例子:

{
  "version": 1,
  "partitions":
     [
        {
            "topic": "Foo",
            "partition": 1         
        },
        {
            "topic": "Bar",
            "partition": 0         
        }
     ]            
}

2.13　刪除topics

[zk: localhost:2181(CONNECTED) 18] get /admin/delete_topics

Schema:
{ "fields":
    [ {"name": "version", "type": "int", "doc": "version id"},
      {"name": "topics",
       "type": { "type": "array", "items": "string", "doc": "an array of topics to be deleted"}
      } ]
}

例子:
{
  "version": 1,
  "topics": ["foo", "bar"]
}

2.14　Topic配置

/config/topics/[topic_name]

[zk: localhost:2181(CONNECTED) 19] get /config/topics/topic2